There's a contradiction in most outbound advice that nobody bothers to resolve.
Cold email guides say two to four emails, no more. Sales cadence guides say eight to twelve touches. Both are widely repeated, and they look like they disagree.
They don't. One is counting emails, the other is counting touches, and the difference between those two numbers is the entire argument for multi-channel outreach.
Here's the data that settles it. In a study of 30+ B2B outbound campaigns run in early 2026, reply rates climbed sharply through touch four, then fell off a cliff.
Touch one produced about 25% of total replies. Touches two and three added 20 to 25% each. Past touch eight on a single channel, you're mostly producing noise.
But that's single-channel data, and this is the part that matters.
A four-touch email sequence followed by two LinkedIn touches and two phone calls produces 50 to 80% more meetings than eight emails alone, on the same list.
Same touch count. Same prospects. The only variable is channel diversity.
What multi-channel outbound actually means
Multi-channel outbound is a coordinated sequence that reaches the same prospect across email, LinkedIn and phone, with each channel doing the specific job it does best.
The word doing the work there is coordinated.
Most teams think they're running multi-channel because they send a cold email, wait, then fire a LinkedIn request a week later with no message attached.
That's not coordination. That's two campaigns annoying the same person.
The prospect shouldn't experience three campaigns. They should experience one person who keeps showing up in sensible places.
Why it works
Three mechanisms, and only one of them is "more touches."
Exposure builds credibility. A single cold email from an unknown sender is indistinguishable from spam. The same message from someone whose LinkedIn profile they saw two days ago reads completely differently.
Channels have different failure modes. Email dies in a spam filter. LinkedIn requests sit unaccepted. Calls hit voicemail.
Running one channel means one point of failure between you and a prospect who might genuinely want what you sell.
Buying committees live in different places. With six to ten people typically involved in a B2B decision, your champion might be an email person while the economic buyer only ever answers a phone.
Channel economics, and what they dictate
| Channel | Cost per touch | Realistic daily capacity | Best job |
|---|---|---|---|
| Fractions of a cent | Thousands, across mailboxes | Wide, cheap filtering | |
| Low, but seat-limited | 15-20 requests per account | Warm touch, plus social signals | |
| Phone | High, rep hours | 50-80 dials | Converting interest into a meeting |
Two numbers from that table decide your sequencing.
Cold call connect rates average around 5.4%, rising to 13.3% for top performers, and it takes roughly eight dials to reach one decision-maker.
But when a rep does connect, conversation-to-meeting conversion reaches 16.7% for top performers against a 4.6% average.
So the phone isn't a bad channel. It's an expensive channel with an excellent conversion rate, which means it should be pointed at your highest-probability conversations rather than your lowest.
The sequencing order that works
1. EMAIL wide, cheap, asynchronous → the whole verified list
│
▼
opens, replies, or accepted connection
2. LinkedIn warm, rate-limited → humanises you, builds recognition
PHONE expensive, high-conversion → only where there's engagement
│
▼
3. Booked meeting
The principle is simple. Lead with email because most contacts open work email. Follow with LinkedIn to humanise the outreach. Close with phone at the highest-intent moment.
And the inverse is worth stating plainly, because plenty of teams get this backwards.
Inverting the order and leading with phone typically reduces total cadence performance, because a phone interruption without prior context feels intrusive.
There's one legitimate exception. When your addressable market is genuinely tiny, a few hundred named accounts, email volume can't carry the programme.
Below roughly 500 contactable prospects the economics flip, and a high-touch, research-heavy motion is correct.
Email-led calling
This is the practical version of the sequencing rule, and it's the highest-leverage change most sales teams can make.
Instead of dialing a cold list, reps only call prospects who already engaged with an email.
Email does the wide, cheap, asynchronous filtering. The phone does what it's genuinely best at, which is turning stated interest into a calendar event.
The rep's call stops being an interruption and becomes a follow-up to something the prospect did yesterday.
We ran this with a mid-sized services client whose team of thirteen reps had been working cold lists.
Switching them to calling only email hand-raisers cut acquisition cost by around 40% and added 39 booked calls the following month. Same reps, same product, same market.
What each channel should actually do
Four touches maximum. Fifty to seventy-five words for the first, thirty-five to fifty for follow-ups.
One thing changed here that most cadence advice hasn't caught up with. Since Microsoft began enforcing bulk-sender requirements in May 2025, high-frequency email-heavy sequences became a liability rather than a strategy.
Eight emails to the same person is no longer just annoying. It's a deliverability risk that costs you the domain.
That's another reason the touch budget should be spread across channels rather than concentrated in the inbox.
Two touches per sequence is the right baseline. A connection request early, then a direct message mid-sequence once they've accepted.
The stay-safe limits matter more than most people realise. Roughly 15 to 20 connection requests a day, 80 to 100 a week, fewer on new accounts.
Cross those and you get restricted, which costs you the account plus any warm relationships inside it. Volume comes from adding seats, never from pushing one account harder.
The tactic worth stealing: LinkedIn voice notes under 60 seconds. They produce 25 to 45% higher response rates than text-only messages.
The reason is partly that they're still rare, and partly that a 40-second voice note takes less effort to absorb than a paragraph of text.
Both advantages decay as adoption rises, which is an argument for using it now.
LinkedIn also earns its place as a sourcing layer before it's a sending layer. Competitor page followers, engagers on a keyword-matched post, people who changed roles last month.
Those are observable signals no firmographic database contains.
One caveat: a follower list is not a buyer list. Qualify against your ICP first, because a large share of any company's followers are job applicants.
Phone
Fifteen to twenty-five percent of your total touches. Not more, because the cost per touch doesn't justify it, and not zero, because it's where meetings actually get agreed.
Timing beats volume here. Morning calls between 7:30 and 9:30 local time, before the calendar fills, or 5:00 to 6:30 in the evening after meetings end.
Give reps the email thread before they dial, so the opener references what the prospect actually did rather than starting cold.
A cadence you can copy
Fourteen to twenty-one days, eight to ten touches, front-loaded in the first three days.
| Day | Channel | Touch |
|---|---|---|
| 1 | Signal-led opener, soft ask | |
| 2 | Connection request with context, no pitch | |
| 4 | Threaded follow-up, new angle | |
| 6 | Phone | Only if there's been engagement |
| 8 | Voice note, if connected | |
| 11 | Fresh thread, different framing | |
| 14 | Phone | Second attempt on engaged contacts only |
| 18 | Clean close that gives something |
Why front-load? Cadences shorter than fourteen days miss the majority of prospects who respond after touch five. Cadences longer than thirty days produce diminishing returns and start burning the relationship.
Front-loading the first three days establishes presence quickly, then the spacing lets it breathe.
Scale length to market size. Very large markets justify fewer touches across more people. Markets under 3,000 prospects justify more, with gaps that stretch, because every contact is irreplaceable.
Fast-track on signals
The best cadence adjustment costs nothing and almost nobody does it.
If a prospect opens an email three or more times, or clicks an asset link, move the next call step up immediately. Don't wait for the scheduled day.
Same for high-intent behaviour like a website visit or a pricing page view. The scheduled cadence is a default for prospects who haven't told you anything.
The moment they do, the schedule should bend around them.
What to measure
Track per channel and in aggregate, but make decisions on the aggregate.
| Metric | Healthy |
|---|---|
| Sequence reply rate | 8-15% for cold 1:1, higher for ABM |
| Meetings booked per sequence | 0.8-2.5% cold 1:many, 4-8% cold 1:1 |
| Meetings from trigger-event outreach | 8-14% |
| Phone connect rate | 8-14% top quartile |
| Conversation to meeting | 4.6% average, 16.7% top performers |
| LinkedIn acceptance | Track the trend, not the number |
That third row deserves attention. Trigger-event outreach books meetings at 8 to 14%, against 0.8 to 2.5% for generic one-to-many sequences.
Which is a reminder that channel mix is a multiplier on targeting, not a substitute for it. A brilliantly sequenced cadence into a firmographic list still loses to a single well-timed email.
Resist single-channel attribution. If someone got an email, accepted a connection, then booked off a call, the call didn't source that meeting. The sequence did.
Teams that attribute at channel level reliably end up cutting the cheap channel that was doing the filtering.
Five ways this goes wrong
1. Opening with the phone. Your most expensive channel pointed at your least-qualified contacts.
2. Running channels in parallel but uncoordinated. An email team and a calling team hitting the same list independently isn't multi-channel. Sequences have to share state.
3. Automating LinkedIn to the limits. Restricted accounts are expensive and slow to replace, and the ceiling arrives faster than the payoff.
4. Repeating the same message across channels. If your LinkedIn message is your email with the greeting changed, you added a touch without adding information.
5. No stop condition. When a prospect replies, every other channel and every colleague at that company should stop immediately.
Company-level auto-pause is a setting in most sequencers. Turn it on for SMB targets, leave it off only for enterprise accounts where multiple separate buyers are legitimate.
Frequently Asked Questions
Does multi-channel really beat email alone?
Yes, consistently. Adding LinkedIn and phone touches to an email sequence lifts meetings booked by 40 to 60%. Teams running properly coordinated multi-touch sequences book two to three times more meetings than single-channel teams. The gain comes from exposure and from covering each channel's failure modes, not simply from more messages.
How many touches should a cadence have?
Eight to twelve across fourteen to twenty-one days, with no more than four of those being emails. Front-load the first three days, then space the rest.
Which channel should I add first?
LinkedIn, in most cases. It's cheaper than phone, it doubles as a sourcing layer for signals no database contains, and it stacks cleanly alongside an email sequence. Add phone once you have enough engaged prospects to keep a rep busy calling people who've already shown interest.
How many LinkedIn connection requests can I safely send?
Roughly 15 to 20 a day and 80 to 100 a week per account, lower on accounts under three months old or without genuine activity. Randomise timing, and scale by adding seats rather than pushing one account harder.
What's the best time to cold call?
Early morning between 7:30 and 9:30 local time, before the calendar fills, or late afternoon from 5:00 to 6:30 once meetings have ended. Both windows work for the same reason. You're catching people between commitments rather than during them.
Should the same person handle all channels for a prospect?
Ideally yes for mid-market and enterprise, where the relationship matters and context gets lost in handoffs. For high-volume SMB motions, splitting sending from reply handling is fine, as long as the reply handler has the full thread and responds inside the day.
How do I stop this feeling like harassment?
Three rules. Every touch carries new information rather than repeating the last. Every channel stops the moment someone replies anywhere. And every sequence has a defined end with a clean close. Prospects tolerate persistence that's useful. They report persistence that's repetitive.
Do voice notes actually work, or is that a fad?
They work now, and part of why they work is that they're still uncommon. Response rates run 25 to 45% above text-only LinkedIn messages. Expect that edge to compress as more people adopt it, which is a reason to use it while it's still novel rather than a reason to dismiss it.
What this comes down to
The instinct when outbound underperforms is to send more. The data says send differently.
Eight emails and eight coordinated touches across three channels cost roughly the same and produce completely different results.
The second gives a prospect several chances to notice you, in contexts where noticing is easy.
Get the order right, respect the limits on each channel, and let signals override the schedule. That's most of the job.
